Temperature error? EThereum miner GTX1070

[SidETH]

I've been running my 6x GTX1070 ethereum miner since July without a problem. This morning the miner software shut down and there was a very light burned smell in the room. I checked everything, but found nothing. Restarted and it was running solid again at 175mhs. After a couple of hours it shut down again. When I now restart it, it runs for a couple of minutes and then starts throwing these temperature errors. It then runs on 3 of the 6 cards. The room is colder than ever with the winterseason coming. Afterburner shows max 55 celcius on the cards.

Suggestions? Am I looking at software or hardware failure?

[P00P135]

check all your risers and power connections for melted cables.

[freezzeer]

I think it's a hardware failure..plss re check your system.

[SidETH]

Well sh*t
Can't see anything obvious, will take out the powersupply tomorrow and see of there are any melted connectors. It burned out 2 months ago (so not so solid running after all), but that was an easy fix under warrenty.

http://www.fotothing.com/photos/ae9/ae9dbd8b92c2bc5d0a4b1ca711622a69.jpg

[SidETH]

One of the SATA power cables to the risers burned at the powerunit (bequiet 1000W platinum). Amazing how something can work perfect for months and then decide to melt.
I have 5 SATA power connectors for 6 videocards. Suggestions on how to connect?

[P00P135]

yeah don't use SATA to power anything, only use molex, 6 and 8 pins.

[SidETH]

Why do we use connectors at all? Why not solder the wires together and skip a connector?

[Cereberus]

One of the SATA power cables to the risers burned at the powerunit (bequiet 1000W platinum). Amazing how something can work perfect for months and then decide to melt.
I have 5 SATA power connectors for 6 videocards. Suggestions on how to connect?

Molex is the best connection to a riser. SATA won't hold for long, you were lucky it run for months without problems, but that was to be predicted when connecting SATA cables to the risers. Buy additional molex cables at ebay or amazon as I am sure your power supply (which should be a fully modular one) will support them easily. All my mining rig which most of them have the RX 580 card(same watt consumption as a GTX 1070) are running since almost one year without problems and are all connected with 4 pin molex to the riser from the PSU (A Corsair Rm 1000i fully modular).

[ROFLED]

Unfortunately, the burning of wires does not depend on the CPU temperature

[Marvell2]

1070 solo mining ahould be 181 mhs

[SidETH]

The thing is it melts the connector at the power unit. So the whole powerunit becomes useless. Yesterday I picked up a new one under warrenty. So, no burning or melting at the SATA connector, but at the drives connector at the power unit...

Connectors: https://content.hwigroup.net/images/products_xl/302730/2/be-quiet-dark-power-pro-11-1000w.jpg

[rednoW]

1070 solo mining ahould be 181 mhs

eth mining speed is the same no matter if you mining solo or on the pool.

[SidETH]

This turns out to be somewhat of a challenge.
Bequiet has these 'nice' modular power drives connectors, 5-pin. But I can't find any cables at all going from modular 5-pin to 6-pin riser, or even 4-pin molex.

[Za1n]

Why do we use connectors at all? Why not solder the wires together and skip a connector?

I have actually done this quite a few times, especially for any risers I would need to otherwise power via Molex cables. I use the butt splice connectors that have solder already in them and are also prewrapped in heat-shrink tubing. A quick crimp and pass over with the heat gun and they are ready to go. You still need to adhere to the no more than two risers per run though, as usually the wires and PSU end connectors are still going to be limited in their current carrying capacity.

[Agozyen]

I've been running my 6x GTX1070 ethereum miner since July without a problem. This morning the miner software shut down and there was a very light burned smell in the room. I checked everything, but found nothing. Restarted and it was running solid again at 175mhs. After a couple of hours it shut down again. When I now restart it, it runs for a couple of minutes and then starts throwing these temperature errors. It then runs on 3 of the 6 cards. The room is colder than ever with the winterseason coming. Afterburner shows max 55 celcius on the cards.

Suggestions? Am I looking at software or hardware failure?

Smell the air vents on your power supply.  If it's shutting down then it's possible that the power supply is starting to fail.  Check to make sure all GPU and PSU fans are running as well.  If you can't find anything wrong and all the cables are fine and so on, set the temperature thresholds lower and underclock the cards.

[SidETH]

Thanks for all the help!
Will work on it tomorrow, hope to get it running again.

[Marvell2]

1070 solo mining ahould be 181 mhs

eth mining speed is the same no matter if you mining solo or on the pool.

Im saying a six gpu 1070 rig is 30 mhs per card or 180

i have 10 differnt brands and they all mine at 30 mhs unless im dual mining

learn to read

[rednoW]

1070 solo mining ahould be 181 mhs

eth mining speed is the same no matter if you mining solo or on the pool.

Im saying a six gpu 1070 rig is 30 mhs per card or 180

i have 10 differnt brands and they all mine at 30 mhs unless im dual mining

learn to read

learn terminology!
Solo mining is mining on your own wallet running on your local network - without using a pool.
ETH mining, or if you want to separate it from dual-mode - use word single mining.

[Kumic]

1070 solo mining ahould be 181 mhs

or 192mh ...or with right 1070 card even higher

[SidETH]

Got the thing up and running again.

However, there are only 3 'drives' connectors available to power 6 risers.
At the moment again over SATA, because I just can't f-in find the correct cables

I have 3 of these available on the modular powerunit to power 6 risers:
http://www.fotothing.com/photos/db7/db7e3f124e82d401fda051454605a541.jpg

And the risers are 6-pin:
https://c.76.my/Malaysia/riser-pci-e-1x-to-16x-sata-usb-3-0-6pin-gpu-riser-pcie-card-60cm-lala55-1706-19-lala55@1.jpg

Any ideas on where to pick up the correct cables?